[r] Add pbmc3k dataset helper (#792)

* Add method for creating seurat pbmc3k dataset
* Update news
* Bump version and update docs

apis/r/DESCRIPTION

@@ -7,7 +7,7 @@ Description: Interface for working with 'TileDB'-based Stack of Matrices,
  from and export to in-memory formats used by popular toolchains like
  'Seurat', 'Bioconductor', and even 'AnnData' using the companion Python
  package.
-Version: 0.1.22
+Version: 0.1.22.9000
 Authors@R: c(
  person(given = "Aaron",
  family = "Wolen",
@@ -45,7 +45,7 @@ Imports:
  urltools,
  vctrs
 Roxygen: list(markdown = TRUE)
-RoxygenNote: 7.2.1
+RoxygenNote: 7.2.3
 Suggests:
  rmarkdown,
  knitr,

apis/r/NAMESPACE

@@ -16,9 +16,11 @@ export(SOMACollection)
 export(TileDBArray)
 export(TileDBGroup)
 export(TileDBObject)
+export(dataset_seurat_pbmc3k)
 import(tiledb)
 importFrom(Matrix,mat2triplet)
 importFrom(Matrix,nnzero)
+importFrom(Matrix,readMM)
 importFrom(Matrix,sparseMatrix)
 importFrom(R6,R6Class)
 importFrom(SeuratObject,AddMetaData)
@@ -37,4 +39,7 @@ importFrom(glue,glue_collapse)
 importFrom(methods,slot)
 importFrom(urltools,url_compose)
 importFrom(urltools,url_parse)
+importFrom(utils,download.file)
 importFrom(utils,modifyList)
+importFrom(utils,read.table)
+importFrom(utils,untar)

apis/r/NEWS.md

@@ -1,3 +1,9 @@
+# tiledbsoma (development version)
+
+## Features
+
+* New function `dataset_seurat_pbmc3k()` to download the pbmc3k dataset from 10X and import as a `Seurat` object without requiring any extra dependencies.
+
 # tiledbsoma 0.1.19
 
 ## Changes

apis/r/R/datasets.R

@@ -0,0 +1,46 @@
+#' Seurat 3k PBMCs from 10x Genomics
+#'
+#' Create a [`SeuratObject::Seurat`] object containing the widely used 3k PBMCs
+#' dataset from 10x Genomics.
+#'
+#' @returns a [`SeuratObject::Seurat`] object
+#' @seealso https://support.10xgenomics.com/single-cell-gene-expression/datasets/1.1.0/pbmc3k
+#' @importFrom Matrix readMM
+#' @importFrom utils download.file read.table untar
+#' @export
+
+dataset_seurat_pbmc3k <- function() {
+ url <- "https://cf.10xgenomics.com/samples/cell-exp/1.1.0/pbmc3k/pbmc3k_filtered_gene_bc_matrices.tar.gz"
+
+ tarfile <- file.path(tempdir(), basename(url))
+ tardir <- sub("\\.tar\\.gz$", "", tarfile)
+ datadir <- file.path(tardir, "filtered_gene_bc_matrices", "hg19")
+
+ if (!dir.exists(tardir)) {
+ if (!file.exists(tarfile)) {
+ utils::download.file(url = url, destfile = tarfile)
+ }
+ utils::untar(tarfile, exdir = tardir)
+ }
+
+ mat <- Matrix::readMM(file.path(datadir, "matrix.mtx"))
+ genes <- utils::read.table(
+ file = file.path(datadir, "genes.tsv"),
+ header = FALSE,
+ col.names = c("id", "gene_name")
+ )
+ barcodes <- utils::read.table(
+ file = file.path(datadir, "barcodes.tsv"),
+ header = FALSE,
+ col.names = "id"
+ )
+ dimnames(mat) <- list(genes$id, barcodes$id)
+
+ object <- SeuratObject::CreateSeuratObject(counts = mat)
+ object[["RNA"]] <- SeuratObject::AddMetaData(
+ object = object[["RNA"]],
+ metadata = genes$gene_name,
+ col.name = "gene_name"
+ )
+ object
+}